Geography Reference
In-Depth Information
rely on a small subset commands, such as picking and highlighting. Consequently, users are
missing out on rich interaction techniques.
Navigation is still tricky. Users get lost in their exploration, they forget where they are and
they may start to misinterpret information. Researchers need to create interfaces that clearly
demonstrate where a user is, what orientation the data is and what scale it is displayed at.
Furthermore, it is important to generate usable and intuitive exploration environments. If
the user does not know how to control the environment, or if they get overwhelmed by too
many windows, then they will not be effective in their exploration process. Thus, developers
need to strike a balance between comprehensive function-rich environments on the one
hand, and simple to use interfaces on the other. Developers should continue to research into
methods that aid the user in their exploration and store session information.
Finally, coordination is obviously a major part in CMV systems. However, coordination
tends to be achieved automatically and often performed silently by a system (such that users
do not necessarily know what is connected together). Hence, meta-visualization techniques
should be developed to enable the user to see how the system is setup.
3.5 Tools and toolkits
Geographical and spatial data has been long used to demonstrate research ideas in coor-
dination. Felger and Schroeder (1992) in the visualization input pipeline (VIP) describe
linked cursors of three-dimensional maps. LinkWinds (Jacobson, Berkin and Orton, 1994)
demonstrates coordination using precipitation and ozone depletion; Visage (Roth
et al.
,
1996) demonstrates coordinated manipulation of maps; the tight coupling interface of
DEVise (Livny
et al.
, 1997) has been used to demonstrate various examples including look-
ing at product purchases by location; Spotfire (Ahlberg, 1996) uses map visualizations along
with other statistical forms.
There are a few different ways to develop CMV systems. Tools such as CommonGIS
(Andrienko and Andrienko, 1999), GeoVISTA studio (Takatuska and Gahegan, 2002) and
Improvise (Weaver, 2004) contain comprehensive coordination capabilities. However, if de-
velopers wish to develop from scratch, then languages such as Java provide a convenient
medium for development. For instance, tools like Mondrian (Theus, 2002) have been de-
velopedinJava.
In addition, toolkits such as the InfoVis toolkit (Fekete, 2004), Prefuse (http://prefuse.
org/) and Piccolo (Bederson, Grosjean and Meyer, 2004) provide developers with the func-
tionality to create complex visualization environments in Java. However, it still takes a lot
of careful planning design and time to create effective geovisualization tools, and often
developers wish to integrate algorithms and tools from different researchers.
One way to develop prototype tools is to use Flash or scalable vector graphics (SVG),
such as those used by Steiner, MacEachren and Guo (2002) and Marsh, Dykes and Atti-
lakou (2006), respectively. These solutions have the added advantage that they are instantly
web accessible and available by many remote users. Alternatively, techniques such as the
Information Visualization CyberInfrastructure (IVC) software framework, which extends
original work on the Information Visualization Repository (Borner and Zhou, 2001), pro-
vide a uniform method to interact with a multitude of algorithms by providing a program-
ming interface to algorithms and a user interface to end-users using the Eclipse Rich Client
Search WWH ::
Custom Search